针对质量控制及诊断过程中出现的不确定、模糊性问题,提出了一种基于模糊逻辑的推理系统.该系统由六个模糊推理子模块及一个聚合模块组成.每个子模块以控制图异常模式数据作为输入,通过模糊化确定异常模式发生的程度,用重心法解模糊后的输出,作为相应异因对输入异常模式的贡献度.不同子模块的输出结果采用模糊算子进行聚合,根据给定阈值对聚合结果排序,形成过程失控情况下异因查找的一个优先序列,并以优先级最高的异因作为排查对象,缩小查找范围,加速异常排除.最后,通过实例表明,该系统町以发现Minitab软件无法预警的质量异常,并可给出相应的排查方案.%Aiming at the uncertain and vague problems during quality control and diagnosis process, a Fuzzy Inference System(FIS) was developed based on fuzzy logic. This system was composed of six fuzzy inference sub-modules and one aggregation module. Each module took the control charts' abnormity pattern data as input. Abnormity pattern presence degree was determined through input fuzzification. By defuzzifying the inference result through centroid method, the contribution intensity of corresponding cause to the input abnormity was obtained in each module. All rnthe evidence supporting each cause from all the modules was aggregated by using fuzzy connective operators, and the final aggregation results were sequenced based on the given fuzzy threshold. The assignable causes search could begin with the cause with the highest priority when process went out of control. Finally, an application example was given. By comparing to the quality software Minitab for the same process data, the application result illustrated that the system could detect the abnormity which Minitab couldn't. Corresponding examining suggestions were also presented.
展开▼
